transformGamPoi

transformGamPoi provides variance-stabilizing transformations for heteroskedastic count data, particularly for Gamma-Poisson models.

Key Features:

• Delta method transformations: Includes transformations such as 'acosh' and 'log(x+1)' based on the delta method.
• Model residual based transformations: Provides Pearson residuals and randomized quantile residuals as variance-stabilizing alternatives.

Scientific Applications:

• Heteroskedastic count data analysis: Stabilizes variance in datasets where variance is not constant, such as sequencing count data.
• Gamma-Poisson model analysis: Supports transformation of data following Gamma-Poisson distributions for downstream statistical analysis.

Methodology:

The package implements two classes of variance-stabilizing transformations: delta method-based approaches and model residual-based approaches using Pearson and randomized quantile residuals.